Task-oriented dialog modeling and action determination

Abstract

An example operation includes one or more of receiving an utterance from a user which is associated with a task, determining, via a machine learning model, that not enough information is available to determine an action based on the utterance, initiating, via a virtual assistant, a dialog with the user and receiving one or more additional utterances from the user via the input device, determining, via the machine learning model, that enough information is available for accomplishing the action based on the one or more additional utterances, determining, via the machine learning model, a task to be performed based on the utterance and the one or more additional utterances, and transmitting an instruction to a system to perform the task.

What is claimed is: 
     
         1 . An apparatus comprising:
 a receiver configured to receive an utterance from a user; and   a processor configured to
 determine, via a machine learning model, that not enough information is available to predict an action based on the utterance, 
 initiate, via a virtual assistant, a dialog with the user and receiving one or more additional utterances from the user via the receiver, 
 determine, via the machine learning model, that enough information is available to determine the action based on the one or more additional utterances, 
 predict, via the machine learning model, a task to be performed based on the utterance and the one or more additional utterances, and 
 transmit an instruction to a system to perform the task. 
   
     
     
         2 . The apparatus of  claim 1 , wherein the processor is configured to determine, via the machine learning model, to prompt the user for an additional utterance based on content included in a most recently-received utterance. 
     
     
         3 . The apparatus of  claim 1 , wherein the virtual assistant is embedded within a navigation system of a transport, the user is a passenger within the transport, and the processor is configured to determine a task of the transport based on the utterance and the one or more additional utterances. 
     
     
         4 . The apparatus of  claim 1 , wherein the processor is configured to determine, via the machine learning model, to request the user to confirm content included in a previously-received utterance based on the machine learning model. 
     
     
         5 . The apparatus of  claim 1 , wherein the processor is configured to predict an intent of the user after each utterance based on an aggregation of utterances during a conversation with the user, and determine that enough information is available based on the aggregation of utterances. 
     
     
         6 . The apparatus of  claim 1 , wherein the processor is configured to determine, via the machine learning model, that a sequences of utterances will be required to make a prediction based on the received utterance. 
     
     
         7 . The apparatus of  claim 1 , wherein the method further comprises predicting, via the machine learning model, a next action to be taken by the virtual assistant based on the received utterance associated and one or more previously-received utterances from the user which are associated with a same task. 
     
     
         8 . The apparatus of  claim 1 , wherein the method further comprises identifying, via the machine learning model, a word of interest within an additional utterance received from the user and an additional question for the virtual assistant to ask the user based on the identified word of interest. 
     
     
         9 . A method comprising:
 receiving an utterance from a user;   determining, via a machine learning model, that not enough information is available to determine an action based on the utterance;   initiating, via a virtual assistant, a dialog with the user and receiving one or more additional utterances from the user via the input device;   determining, via the machine learning model, that enough information is available to determine the action based on the one or more additional utterances;   determining, via the machine learning model, a task to be performed based on the utterance and the one or more additional utterances; and   transmitting an instruction to a system to perform the task.   
     
     
         10 . The method of  claim 9 , wherein the determining that not enough information is available comprises determining, via the machine learning model, to prompt the user for an additional utterance based on content included in a most recently-received utterance. 
     
     
         11 . The method of  claim 9 , wherein the virtual assistant is embedded within a navigation system of a transport, the user is a passenger within the transport, and the determining the task comprises determining a task of the transport based on the utterance and the one or more additional utterances. 
     
     
         12 . The method of  claim 9 , wherein the determining that not enough information is available via the machine learning model comprises determining to request the user to confirm content included in a previously-received utterance based on the machine learning model. 
     
     
         13 . The method of  claim 9 , wherein the determining that enough information is available via the machine learning model comprises predicting an intent of the user after each utterance based on an aggregation of utterances during a conversation with the user, and determining that enough information is available based on the aggregation of utterances. 
     
     
         14 . The method of  claim 9 , wherein the method further comprises predicting, via the machine learning model, that a sequences of utterances will be required to make a prediction based on the received utterance. 
     
     
         15 . The method of  claim 9 , wherein the method further comprises predicting, via the machine learning model, a next action to be taken by the virtual assistant based on the received utterance associated and one or more previously-received utterances from the user which are associated with a same task. 
     
     
         16 . The method of  claim 9 , wherein the method further comprises identifying, via the machine learning model, a word of interest within an additional utterance received from the user and an additional question for the virtual assistant to ask the user based on the identified word of interest. 
     
     
         17 . A computer-readable storage medium comprising instructions, that when read by a processor, cause the processor to perform a method comprising:
 receiving an utterance from a user;   determining, via a machine learning model, that not enough information is available to determine an action based on the utterance;   initiating, via a virtual assistant, a dialog with the user and receiving one or more additional utterances from the user via the input device;   determining, via the machine learning model, that enough information is available for accomplishing the action based on the one or more additional utterances;   determining, via the machine learning model, a task to be performed based on the utterance and the one or more additional utterances; and   transmitting an instruction to a system to perform the task.   
     
     
         18 . The computer-readable storage medium of  claim 17 , wherein the determining that not enough information is available comprises determining, via the machine learning model, to prompt the user for an additional utterance based on content included in a most recently-received utterance. 
     
     
         19 . The computer-readable storage medium of  claim 17 , wherein the virtual assistant is embedded within a navigation system of a transport, the user is a passenger within the transport, and the determining the task comprises determining a task of the transport based on the utterance and the one or more additional utterances. 
     
     
         20 . The computer-readable storage medium of  claim 17 , wherein the determining that not enough information is available via the machine learning model comprises determining to request the user to confirm content included in a most-recently received utterance based on the machine learning model.